Wherever you are in the world it is unlikely you will have missed the reporting of last week’s events at UK airports, in particular London’s Heathrow. In a nutshell the UK Government announced, in response to highly creditable police work in detaining alleged terrorists, the very highest level of airport security. The result was nothing short of chaos.

I note Grady Booch, interviewed recently has made some interesting comments on SOA including: "It is simply a mechanism for reaching into systems. You see organizations rushing to [implement] services, but they are really missing the fundamental engineering principles. In about 18 months, they will complain SOA doesn’t work. They’ll be blaming the wrong thing. They should be blaming their architectures and best practices."

My take on this is that we are moving further away from the old-fashioned idea of a single all-powerful general-purpose programming language/environment. The complexity (and sometime popularity) of PL/1 and ADA resulted from the expectation that you could (and should) code everything you ever wanted in a single syntax.

I note Microsoft has completed the acquisition of Softricity, a provider of application virtualization and dynamic streaming technologies. Softricity provides Microsoft with an application and OS management process that will create a foundation for software services infrastructure. Using Softricity's application virtualization technology  we can expect applications will be served centrally and delivered directly to the user's desktop in an isolated, virtualized image, minimizing application-related alterations to the operating system and compatibility challenges with other applications.

Combined with Softricity's software streaming functionality, virtualized applications will be dynamically sent over the network, effectively whenever and wherever a user needs them. The approach creates the option of a centrally managed software service in which applications are available when and where needed and are easily updated.

Every so often an enterprising industry analyst manages to create a huge debate by challenging some of the pillars of the industry. The debate de jour was sparked off by the Burton Group suggesting the Java Platform, Enterprise Edition (Java EE) is not going to survive as a major standard programming model in the next five years. Like one of those prehistoric animals that went extinct because it got too big to live off the available foliage, the Burton analyst said that with the release this spring of JEE5, the Java EE platform has grown too complex to be workable for enterprise developers, who are increasingly looking at alternatives such as Ruby-on-Rails.

It seems that there is in general a strong trend to the casual, informal and unstructured, epitomized by the contrast between Microsoft and Google. However the danger is that we assume all classes of problem can be solved in the same way. Real world enterprise class problems usually contain a mix of high and low structure, and require an architectural framework and policy structure that reflects this. 

I have been quite concerned recently that many commentators and advisors who should know better have been characterizing SOA as instant gratification. In the last few weeks I attended a major conference in Washington DC where supposedly influential people were advising – “don’t worry about delivering the right services - because they are inherently flexible they can be changed, orchestrated or replaced so easily”. Another speaker suggested that “the answer to today’s explosion in services was to use semantic ontologies to discover and patch together services that were never intended to cooperate”. 

What is the biggest challenge facing organizations in their adoption of SOA? From countless discussions with them, I conclude it is the organization itself.

Organizations can acquire the necessary SOA Infrastructure from countless vendors of relatively mature technology and products. They can be taught the processes and methods required by consultants such as ourselves. But organizational and cultural changes must come from within.

What makes an organization successful in their adoption of SOA? Clearly we can provide guidance on what we believe organizations should be doing but what are those organizations who have achieved success with SOA already doing? Why are some organizations already delivering benefits measured in millions of dollars attributed to SOA, whilst others are still finding it difficult to get started?

Analysing the various case studies around and our own experience with customers we can identify the following top ten traits of the successful SOA organization

A group of vendors have recently announced SOA Link  which they describe as an “an end-to-end SOA Governance Interoperability Initiative.”

I presented a session on SOA Governance at the recent OMG SOA, MDA and Web Services Workshop. In the presentation I highlighted the challenge of governance in the Service lifecycle when potentially so many different tools are involved in the end-to-end process. For example, the complete set of meta data that specifies the service might be dispersed to several products, across different stages of the lifecycle. 

Recently I reported on how IBM were applying SOA within the Tivoli product set itself to enable Service Oriented Management. Whilst seemingly every IT vendor now encourages the use of SOA, and provide products to support adoption by end-user organizations, there is less evidence that vendors themselves have taken SOA to heart internally.

The Systems Management domain has been the most forward thinking in this respect. For example, CA, BMC, IBM, HP, and others have developed the WSDM specification, and have or are implementing in their respective SM products. The driver for WSDM was not just interoperability between SM products and manageable resources, but enabling federated management that enables distributed resources to be managed across the Internet without requiring proprietary agents at each endpoint.

But WSDM is just one part of the distributed management jigsaw. It defines a protocol by which differnt SM products and resources can monitor status and set the resource configuration for example.

Last week, IBM announced their SOA Governance portfolio of software and professional services designed to help organizations ensure they are achieving their goals and objectives for SOA adoption.

On one hand SOA governance can be considered just part of IT governance. Ensuring that business requirements are met, delivering ROI, and meeting SLA is the same for SOA as any other IT activity. Organizations should not be worrying about putting SOA governance in place if they don’t already have effective IT governance processes.

Last week I presented at the OMG SOA, MDA and Web Services Workshop.

Recently, OMG launched a SOA SIG. The primary goals of the SIG are to

• support an MDA approach to SOA that links architectural, business and technology views of services, including Business Process Management (BPM) and Event-Driven Architecture (EDA).
• Identify and foster development of OMG modeling standards for SOA that integrate with and complement standards developed by other organizations such as W3C, Open Group and OASIS.
• To improve awareness and understanding of SOA by OMG members.
• To coordinate SOA related efforts within OMG.

Right now the SOA SIG cupboard is rather empty, but Pete Rivett  of Adaptive did give a presentation at the meeting covering some of the possible activities that this SIG might focus on.

Werner Vogels (Amazon CTO) has just posted some notes about Scalability, which he defines thus: A service is said to be scalable if when we increase the resources in a system, it results in increased performance in a manner proportional to resources added.

Scalability is relevant for tools and methods. I have been shown demonstrations of tools that look pretty cool when you've got five web services, but are pretty hopless when you've got five hundred or five thousand web services. What does a tool do when you have more objects than can reasonably appear on a screen - even with scroll bars? Was the multi-screen, multi-user option conceived from the start, or does it look like an afterthought?

I also attended a presentation last week that was titled Agile Architecture. The presentation turned out to be a somewhat evangelical presentation on agile methods. But agile architecture it was not. To be fair the presentation did mention “project” architecture, but then rather depreciated the thought by saying that on an agile project everyone is an architect. Really?

When asked about enterprise architecture the presenters just brushed the question off as not really relevant which upset a good few people in the audience – it was a conference about architecture for architects.